Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

$ y_n'' -nx\frac{1}{\sqrt {y_n}}=0$

Is there any known method to solve such second order non-linear differential equation?

What I tried to solve:

$ 2y_n'y_n''= 2nx\frac{y_n'}{\sqrt {y_n}}$

$ y_n'^2= 4nx\sqrt {y_n}-4n\int\sqrt {y_n} dx$

After that I could not see any way how to proceed.

Please advice what to do to solve the differential equation.

Many thanks

share|improve this question
    
from Maple I got: $$ \int^{y(x)} {1 \over \sqrt{4nx \sqrt a - C_1}}da- x - C_2$$ –  Santosh Linkha Aug 29 '12 at 8:44
    
@experimentX : thanks a lot. is the result equal to zero? any idea how to get the result manually? –  Mathlover Aug 29 '12 at 8:51
    
yes it was ... and where were two!! –  Santosh Linkha Aug 29 '12 at 8:54
    
the other was $$\int^{y(x)} -{1 \over \sqrt{4nx \sqrt a - C_1}}da- x - C_2$$ –  Santosh Linkha Aug 29 '12 at 8:55
    
let's try something like this $$ {d^2x \over dy^2} = {\sqrt{y_n} \over nx}$$ –  Santosh Linkha Aug 29 '12 at 9:00

2 Answers 2

up vote 1 down vote accepted

This is indeed a Emden-Fowler equation as found at EqWorld.

Polyanin provides a parametric solution to $\ y''=n\,x\,y^{-1/2}\ $ as $(2.3.1.11)$ in his book 'Exact Solutions for Ordinary Differential Equations' : $$ \begin{align} x(t)&=a\;C_1\,e^{-t}\left(e^{3t}+C_2 \sin(\sqrt{3}t)\right)\\ y(t)&=b\;(C_1)^2\,e^{-2t}\left(2e^{3t}-C_2 \sin(\sqrt{3}t)+\sqrt{3}C_2 \cos(\sqrt{3}t)\right)^2\\ &\quad\text{with}\ \ n=16\,a^{-3}\,b^{\frac 32} \end{align} $$

(you may find this too at Amazon searching $\ $Axy$\ $ and selecting page 309)

In fact there are more constants than necessary and one of the constants $a,\ b,\ C_1$ may be set to $1$ without restricting the solutions. To verify this set $\ b:=d^2\ $ getting $\ x(t)=(a\,C_1)^2\cdots$, $y(t)=(d\,C_1)^2\cdots\ $ with the constraint $n=16\bigl(\frac da\bigr)^3$ so that $d$ and $a$ are defined up to a multiplicative constant even without $C_1$.

To verify that the solution provided 'works' we may evaluate : $$\frac {dy}{dx}=\frac {y'(t)}{x'(t)}=\frac {4 C_1 b}a e^{-t}\left(2e^{3t}-C_2 \sin(\sqrt{3}t)-\sqrt{3}C_2 \cos(\sqrt{3}t)\right)$$ $$\frac {d \frac {dy}{dx}}{dx}=\frac {16 b}{a^2} \frac{e^{3t}+C_2\sin(\sqrt{3}t)}{2e^{3t}-C_2 \sin(\sqrt{3}t)+\sqrt{3}C_2 \cos(\sqrt{3}t} $$ that we may compare with $\dfrac {x(t)}{\sqrt{y(t)}}$ (the ratio should be $\frac {16\,b^{3/2}}{a^3}$)

Of course a detailed proof would be better.

Hoping this helped anyway,

share|improve this answer

In fact it belongs to an Emden-Fowler equation.

First, according to http://eqworld.ipmnet.ru/en/solutions/ode/ode0302.pdf or http://www.ae.illinois.edu/lndvl/Publications/2002_IJND.pdf#page=6 , all Emden-Fowler equations can be transformed into Abel equation of the second kind.

Let $\begin{cases}u=\dfrac{x^3}{y_n^\frac{3}{2}}\\v=\dfrac{x}{y_n}\dfrac{dy_n}{dx}\end{cases}$ ,

Then $\dfrac{dv}{du}=\dfrac{\dfrac{dv}{dx}}{\dfrac{du}{dx}}=\dfrac{\dfrac{x}{y_n}\dfrac{d^2y_n}{dx^2}+\dfrac{1}{y_n}\dfrac{dy_n}{dx}-\dfrac{x}{y_n^2}\left(\dfrac{dy_n}{dx}\right)^2}{\dfrac{3x^2}{y_n^\frac{3}{2}}-\dfrac{3x^3}{2y_n^\frac{5}{2}}\dfrac{dy_n}{dx}}=\dfrac{\dfrac{x}{y_n}\dfrac{d^2y_n}{dx^2}+\dfrac{v}{x}-\dfrac{v^2}{x}}{\dfrac{3u}{x}-\dfrac{3uv}{2x}}=\dfrac{\dfrac{x^2}{y_n}\dfrac{d^2y_n}{dx^2}+v-v^2}{3u\left(1-\dfrac{v}{2}\right)}$

$3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}=\dfrac{x^2}{y_n}\dfrac{d^2y_n}{dx^2}+v-v^2$

$\dfrac{x^2}{y_n}\dfrac{d^2y_n}{dx^2}=3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}+v^2-v$

$\dfrac{d^2y_n}{dx^2}=\dfrac{y_n}{x^2}\left(3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}+v^2-v\right)$

$\therefore\dfrac{y_n}{x^2}\left(3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}+v^2-v\right)-\dfrac{nx}{\sqrt{y_n}}=0$

$\dfrac{y_n}{x^2}\left(3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}+v^2-v\right)=\dfrac{nx}{\sqrt{y_n}}$

$3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}+v^2-v=\dfrac{nx^3}{y_n^\frac{3}{2}}$

$3u\left(1-\dfrac{v}{2}\right)\dfrac{dv}{du}+v^2-v=nu$

$3u\left(\dfrac{v}{2}-1\right)\dfrac{dv}{du}=v^2-v-nu$

Let $w=\dfrac{v}{2}-1$ ,

Then $v=2w+2$

$\dfrac{dv}{du}=2\dfrac{dw}{du}$

$\therefore6uw\dfrac{dw}{du}=(2w+2)^2-(2w+2)-nu$

$6uw\dfrac{dw}{du}=4w^2+6w+2-nu$

$w\dfrac{dw}{du}=\dfrac{2w^2}{3u}+\dfrac{w}{u}+\dfrac{2-nu}{6u}$

In fact, all Abel equation of the second kind can be transformed into Abel equation of the first kind.

Let $w=\dfrac{1}{z}$ ,

Then $\dfrac{dw}{du}=-\dfrac{1}{z^2}\dfrac{dz}{du}$

$\therefore-\dfrac{1}{z^3}\dfrac{dz}{du}=\dfrac{2}{3uz^2}+\dfrac{1}{uz}+\dfrac{2-nu}{6u}$

$\dfrac{dz}{du}=\dfrac{(nu-2)z^3}{6u}-\dfrac{z^2}{u}-\dfrac{2z}{3u}$

Please follow the method in http://www.hindawi.com/journals/ijmms/2011/387429/#sec2.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.